Digital BW, The Print

Paul

My answer was based on simple maths as you say defined by the parameters of
a gamma = 2.2 space such as Gray Gamma 2.2 or even SRGB.  If the monitor and
printer are properly calibrated this mid point should print with a density
of 0.61 and match the screen.  Whether people, when looking at the scale,
would agree that 50% at D = 0.61 or a block of pixels with normalised vales
of 0.5 displayed on a Gray Gamma 2.2 screen are fair descriptions of "middle
grey" is entirely different question.

Ansel defined mid grey at 18% reflectance, decided which part of the scene
he wanted to have 18% reflectance in the final print and then let all the
other parts fall where they lay. I haven't been through it yet but it is
interesting to note on Norman Koren's site (again):

http://www.normankoren.com/zonesystem.html

"The charts below are derived from an equation intended to make zone 5 on a
properly calibrated monitor appear as subjective middle gray‹ about 21% of
the maximum screen brightness level, similar to middle gray on a print (18%
reflectance referenced to 90% white). I include the equation in a box near
the end."

When you pull up a step wedge on screen and run the cursor over it examining
the luminance levels you might show D = 61 on screen.  But when it comes to
printing the Epson driver (without benefit of colorsync) performs a tonal
compression.  If your curve shifts 50% = 0.61 so that it prints fine, I am
not sure that a displayed and curve adjusted 50% step is still at pixel
value 0.5.  (But in my proposed methodology where I have access to the
inside of the RIP I could be assured that norm PV = 0.5 prints at 0.61
because any and all tonal adjustments are in my single PS curve.)

Steve


> From: Paul Roark <paul.roark@...>
> 
> So, the question is, empirically, what midtone 50% print density best
> matches the monitor.
> 
> Your numbers, below, suggest that by at least some theory my current target
> of 0.61 might be "correct."  I also have a theory as to why it is "correct,"
> but I also have a theory that suggests it is too light.  Frankly, I based my
> 0.61 standard on an empirical averaging of my and others' Piezo outputs some
> years ago, where consistency with old files was the main concern.
> Currently, however, I'd say most seem to think that the 0.61 50% target is
> too light.  I assume these people are basically saying that the prints are
> "lighter" appearing than their monitors.  As such they need to go through
> some cumbersome process to match the two.  If that matching were better
> built into the target, it might make life easier for many.
> 
> (Although I was once a math major, my experience has convinced me that
> surveys and other empirical studies result in the best models.)
> 
> Paul
> www.PaulRoark.com
> 
>